11 artment of physics and astr - connecting repositories11 li c. r. chinn dep artment of physics and...
TRANSCRIPT
NU
CL
-TH
-941
0033
Correlations in a Many-Body Calculation of 11Li
C. R. Chinn
Department of Physics and Astronomy
Vanderbilt University, Nashville, TN 37235
J. Decharg�e and J.-F. Berger
Service de Physique et Techniques Nucl�eaires
Centre d'Etudes de Bruy�eres-le-Chatel
B.P. No. 12, 91680 Bruy�eres-le-Chatel, FRANCE
Abstract
A many-body calculation of 11Li is presented where the only input is the well-
tested, �nite-range D1S e�ective interaction of Gogny. Pairing correlations are in-
cluded in a constrained Hartree-Fock-Bogolyubov calculation, while long-range col-
lective correlations are introduced using a GCM derived calculation. Correlations are
found to play an important role in describing 11Li. A substantive underlying 9Li core
of 11Li is found, which has a di�erent density pro�le than a free 9Li nucleus. This
may have signi�cant implications in the use of a three-body framework in studies of
I. INTRODUCTION
With the recent advent of secondary beam facilities there has been a large research
interest in nuclei near the drip lines. These exotic nuclei o�er opportunities to study many-
body e�ects under unusual conditions. An example of such can be related to the nucleon-
nucleon [N-N] interaction. While the free N-N interaction appears to be well understood,
the role of the N-N interaction in microscopic nuclear structure is far from being clear. (One
example where one would think that the role of the N-N interaction should be clear, but is
not is in the 3H problem.) Because nuclei near the drip line have weak binding energies and
hence large density distributions, the N-N interaction can now be studied in regions of low
nuclear density.
The free neutron-neutron [n-n] interaction is attractive, but the di-neutron system is
unbound. Migdal and Watson postulated the possibility that a di-neutron may become
bound if placed within the �eld of a nucleus. 11Li among others appears to be such a
system, since both 11Li and 9Li are bound while 10Li is not. Hence, with this in mind, recent
interest in 11Li has been strongly addressed within a three-body framework [1{6], where 11Li
is represented as two neutrons surrounding an inert 9Li core. For a recent review of such
work, please see Ref. [1].
The assumption that 11Li can be realistically represented as a three-body system needs
to be investigated. While present experimental evidence indicates that this assumption may
be valid [7,8], one would like to test this hypothesis within a many-body theoretical inves-
tigation. Di�culties that may arise with the three-body framework include: 1.) treatment
2
of the Pauli exclusion principle can only be performed approximately, and 2.) correlations
with the 9Li core are ignored.
The most straightforward way to study 11Li in an A-body framework is to use a single
particle model, but this was shown for this nucleus [9] to give an inadequate description.
Therefore, for a proper A-body calculation a correlated description beyond a single particle
model is required. By renormalizing the mean �eld potential several groups [9{11] found
that the experimental two neutron separation energy and the rms radius of 11Li could be
reproduced, thus indicating the possible existence of correlations among the outer single
particle states. An attempt to model correlations in a shell model description [12] resulted
in small e�ects, but as will be shown later in this paper the adopted shell model space used
in this study is clearly too small.
In this paper we introduce long-range correlations by using a Generator Coordinate
Method [GCM] type formalism. The nuclear ground state [GS] is represented as a superpo-
sition of HFB nuclear states, which are obtained by constraining on di�erent values of the
hr2i collective variable. The reason for choosing this particular GCM variable can be under-
stood in the following way: correlations in the 11Li GS are expected to occur because the
loosely bound outer neutrons can occupy a large number of nearly degenerate Rydberg-type
orbits having a broad range of radial extensions. One expects that a constraint imposing
di�erent values of the total rms radius will act essentially on the outer neutrons radial dis-
tributions and therefore will be able to generate the kind of con�gurations present in the
nuclear GS. Therefore, as is usual in GCM calculations, initially, a series of constrained HFB
microscopic mean �eld calculations is performed. In a second step the coe�cients of the
3
GCM con�guration mixing are computed by solving the coupled equations resulting from
the application of a variation principle to the total nucleus binding energy. In the present
work, this is accomplished in an approximate fashion by reducing the usual Hill-Wheeler
equations to a collective Schr�odinger equation of the Bohr type. Note that this calculation
includes pairing correlations as well as long-range collective degrees of freedom as derived
from the e�ective N-N interaction. Also, the Pauli principle is strictly obeyed through the
use of fully antisymmetrized single particle wave functions.
The only input into this consistent A-body calculation is the well-tested D1S Gogny
force. This interaction is a density-dependent phenomenological parametrization of the N-N
interaction inside the nuclear medium which includes a spin-orbit term, and is �nite-ranged.
The parameters of the interaction have been �xed by matching the bulk properties of nuclear
matter and of a few �nite nuclei, including pairing correlation strengths. This D1S force
has been tested in a variety of applications with excellent results [13,14]. It must be noted
that this force gives a very good description not only of medium and heavy nuclei, but also
of very light nuclei. For instance it describes correctly the binding energy and radius of
the alpha particle. In addition the �nite range, density-independent part of the force has
been set-up in order to roughly simulate a free N-N interaction in the sense that it gives
the correct N-N scattering lengths. These properties are important in the present context
which deals with a three-proton system where almost free neutron-neutron interactions are
expected to play a crucial role.
With this collective model a qualitative study of correlations in 11Li is presented. In
particular the role of the 9Li core is explored, especially its relation to a free 9Li nucleus.
4
The validity of the three-body hypothesis is investigated using a collective A-body model.
In Section II the mean �eld constrained and unconstrained HFB calculation is described
and the results presented. Long-range correlations using a simpli�cation of the GCM is
described in Section III along with the results and analysis, followed by a conclusion.
II. CONSTRAINED HFB CALCULATION
The constrained HFB A-body calculation is performed using a nineteen shell axially-
symmetric harmonic oscillator [HO] basis. In performing tests of the convergence of the
basis, it was found that a nineteen shell basis is required due to the large extension of the
neutron matter distribution from the center of the nucleus in coordinate space, and to the
inadequate asymptotic behavior of HO states in r space. The use of a multi-oscillator basis,
i.e. of a basis composed of several sets of concentric HO states associated with di�erent
lengths is also used in the HFB calculation. This kind of basis allows one to extend to larger
distances the radial description of nucleon orbits. However, in this case only spherically
symmetric nuclear distributions could be described. Time-reversal symmetry is assumed,
so the protons are described by blocking with equal weights the two j�z = �3=2� axial
quasi-particle orbitals, thus matching the known ground state spin of both 11Li and 9Li.
For a study of a small nucleus such as 11Li, it would be expected that a mean �eld
description would not be the most appropriate choice. In this case we wish to address certain
many-body questions and to test assumptions about the many-body nature of the problem.
With this in mind as a �rst step, a mean �eld calculation using HFB should be able to
5
address some of these qualitative concerns.
As explained in the Introduction, the 11Li nucleus is expected to be very soft against
changes in the density distribution rms. For this reason a constrained HFB calculation was
performed, where the constraint variable used corresponds to the mean value of hr2i:
q =Zdr r2 �(r) ; where �(r) =
Zd �(~r) ; (1)
hrrmsi =pq
When the constraint is switched o�, one gets the mean-�eld representation of the 11Li
GS. The total rms matter radius obtained in this way is about 2.80 fm, matching the results
of several other groups [12,9]. The separate proton and neutron rms radii are found to be
2.30 and 2.97 fm, respectively. For a similar HFB calculation of the 9Li GS the rms radii
are found to be 2.47, 2.24, and 2.58 fm for the total, proton and neutron distributions,
respectively. The rms radii for the protons in 9Li and 11Li di�er only by a small amount in
these mean-�eld calculations, resulting from the e�ects of the force between the neutrons and
the protons. However, when correlations are introduced in 11Li, if there is an uncorrelated
9Li core in 11Li , it may well di�er from the free 9Li GS. It remains to be seen though how
large this di�erence may be.
In Fig. 1 the results of the constrained 11Li calculation are shown as a function of the
constraint variable, q. A similar spherical 9Li calculation is also shown. As q is increased,
the 11Li curve is much softer than 9Li in the sense that the slope is much less steep. One
then expects 11Li to include signi�cantly more con�guration mixing than 9Li. Such a mixing
6
has been included and will be discussed in detail later.
The rms radii for the neutron and proton distributions are shown in Fig. 2a for the
constrained HFB results as a function of hrrmsitotal =pq. As expected (see the Introduction)
the proton rms is una�ected by the constraint, while the neutron hrrmsin varies linearly. In
other words, the curve con�rms that the potential between the protons and outer neutrons
is not strong enough to prevent the two distributions from decoupling.
The independent nature of the proton and neutron sectors is also evident in the pairing
energy shown in Fig. 2b. The protons consistently have zero pairing as a function of q. The
neutrons have strong pairing for q > 3 fm indicating the onset of a signi�cantly high neutron
level density at the Fermi surface.
Since the protons do not exhibit pairing there must be a sizeable gap at the Fermi
surface. This is shown in Fig. 3a, where the protons occupying the 1s1=2 and 1p3=2 states.
There is a gap of about 6 MeV between the 1p3=2 level and the higher single particle levels
and this gap remains for all q considered. The relative energies of the occupied levels do not
change signi�cantly as a function of q, con�rming the negligible in uence of the constraint
on the proton mean-�eld. Note that, since the 1p3=2 level does not shift a great deal,
the blocking approximation used here to account for the odd number of protons should be
reasonable.
In Fig. 3b the corresponding neutron single particle levels are shown. The lowest six
neutrons are almost completely contained in the 1s1=2 and 1p3=2 orbitals. Due to pairing
correlations the last two neutrons are dispersed throughout the higher levels. There is almost
7
no gap between the 1p1=2 level and the higher levels, and for large q these levels cross. The
levels higher than 2d3=2 are not shown. The level density at the Fermi surface is very high,
indicating that con�guration mixing, i.e. the existence of correlations is widespread.
This result is an indication that the model space required to describe the 11Li GS in
an extended shell model calculation is much larger than generally assumed [12]. According
to the present calculation levels above the f shell surely contribute to the GS description.
In Fig. 3b it is clear that there is a large gap between the six inner neutrons, which
represent a 9Li core, and the outer two valence neutrons. The calculated occupation prob-
abilities for the core neutrons are almost always between 1:00 and 0:99 with a minimum of
�> 0:987. This is true for all q considered. Because of these features, a 9Li core wave function
can be projected out by taking the constrained HFB 11Li solutions and explicitly setting the
�rst six neutron levels to have one occupation probability and the other neutron levels to
be empty. As evidenced by the fact that the core neutron probabilities are not exactly one,
this is an approximate procedure, but clearly, because of the large energy gap, this should
be a very reasonable and accurate representation.
In Fig. 4 the neutron density pro�les are shown for the unconstrained HFB calculations
of 9Li and 11Li, and for the 9Li core projected from 11Li as explained above. At the center of
the nucleus the free 9Li neutron density is sizeably larger than the 9Li core neutron density,
and accordingly, the 9Li core extends somewhat farther. More precisely, the central neutron
density of 9Li (short-dashed curve) is about 15% larger than that of the 9Li core (long dashed
curve). A similar di�erence was found for the protons. This is more easily seen when the
8
tail is expanded in a logarithmic plot. The the 11Li core neutrons density doesn't fall o�
as rapidly beyond r = 7 fm. Already one can see evidence for the beginning of a halo-like
structure.
III. CALCULATION INCLUDING CONFIGURATION MIXING
From the evidence shown in the previous two sections it is clear that a pure single
particle model is inadequate to describe 11Li. To provide a more sophisticated representation,
a correlated ground state wave function is constructed as a superposition of the HFB nuclear
states in the following GCM form:
j�oi =Zdq fo(q) j�qi ; (2)
where j�qi is a product of HFB quasiparticle states for deformation q, fo(q) is a weight
function, and q is the constraint variable. In the GCM formalism a hamiltonian kernel
is constructed with the GCM wave function. By applying the variational principle, an
equation is derived from which the weight functions can be calculated. A Gaussian overlap
approximation [GOA], where the overlap between any two HFB states is approximated by
a Gaussian in the collective variable [15], is applied to simplify these equations, deriving
a Bohr Hamiltonian expression. The solution of the Bohr Hamiltonian equation gives the
weight function, fo(q). These techniques have been thoroughly tested in many instances
[16]. It may be that the GOA is not as accurate for smaller nuclei as in previous experience,
but for the investigative study being performed here, this should be more than adequate.
9
The resulting ground state collective wave function, fo(q), has a corresponding total
binding energy. To properly calculate the two neutron separation energy it is necessary to
perform a similar collective calculation for the free 9Li case and then to take the di�erence
in the total collective binding energies. As shown in Ref. [9] the energies calculated from the
single particle HFB model are inadequate and it is necessary to include more sophisticated
degrees of freedom.
The ground state correlated density is calculated from (2) in the following fashion:
�GS(r) =Zdq0
Zdq fo(q
0) h�q0 j �o(r) j�qi fo(q); (3)
where �o(r) is the radial density operator obtained after angle averaging.
In Fig. 5 the proton density pro�le is shown from the unconstrained HFB calculation
of 9Li and 11Li along with the GCM-type collective result for 11Li. As is apparent here the
9Li and 11Li proton pro�les are not equivalent, implying di�erences between the free 9Li
nucleus and the 9Li core of 11Li. The correlated 11Li proton density is very similar to the
uncorrelated 11Li result, which should not be surprising in light of the unchanging single
particle spectrum in Fig. 3a. For the protons the correlations have little e�ect upon the
density distributions both near the central part and along the tail. It appears clear that
there are few correlations in the 11Li proton sector.
The various neutron distributions are shown in Fig. 6. As in the unconstrained HFB
case an approximate representation of the 9Li core can be projected out from the 11Li calcu-
lation including con�guration mixing. This is performed by setting the neutron occupation
propabilities in the constrained HFB solutions to be either one or zero and then using these
10
density matrices with the weight function, fo(q), obtained in the collective 11Li calculation.
As for the proton case the in uence of correlations on the 9Li core appears quite small. The
collective long-range correlations represented by the GCM therefore have little e�ect on the
9Li core as a whole.
This in uence is much bigger on the full 11Li neutron distribution, that is when one
includes the two extra neutrons. The surface of the neutron distribution of 11Li is at about
� 2:5 fm, which is much further extended than the surface of the proton distribution at
� 1:8 fm. Correlations slightly reduce this di�erence. Substracting the 9Li core neutron
distribution from the 11Li neutrons gives the structure of the two valence neutrons. The two
neutrons have zero density at the origin due to the Pauli blocking from the core and extend
in a halo-like structure. At the surface the collective valence structure causes the collective
calculation to have a di�erent neutron density pro�le than the unconstrained HFB neutrons.
This is most easily seen on the logarithmic plot in the lower part of Fig. 6 : the HFB
and GCM GS 11Li neutron densities strongly di�er beyond q = 6 fm. At this point it must
be emphasized that an unphysical ledge appears in the GCM densities beyond q = 7�8 fm.
This ledge is small in amplitude, but it a�ects greatly the calculated rms radius since it
extends very far out. The origin of this ledge can be traced back to the structure of the HFB
constrained solutions included in the con�guration mixing. In Fig. 7 the neutron densities
obtained in constrained HFB calculations using the nineteen shell basis at q = 3:4 fm and
q = 4 fm are shown. Also shown are the same results using a multi-oscillator spherical HFB
basis. This particular multi-oscillator basis uses three concentric sets of eight shell bases
with three di�erent oscillator lengths. This basis corresponds to a very large single oscillator
11
basis (� 60 shells). Clearly, a ledge appears in the HFB densities which depends on the basis.
With the large multi-oscillator basis the ledge appears two orders of magnitude smaller and
at larger r. This ledge is a result of the constraint on hr2i, that tends to push up the density
at large r, but then is restricted by the local nature of the harmonic oscillator bases. One in
fact observes the parabolic fall of the densities in the 19 shell HO basis, characteristic of the
HO asymptotic behavior at large r. A similar parabolic fall of the multi-oscillator densities
is also observed at very large r (30 fm). Both the property of r2 to be unbounded and the
use of a restricted HO space may be responsible for this phenomenon
Since there clearly exists a great deal of con�guration mixing in 11Li due to the radial
extension of outer neutrons, one would also like to be able to extract a reasonable asymptotic
tail for the GS wave function. This is necessary to get a reasonable estimate of the GS neu-
tron rms radius. The single particle wave functions should asymptotically be proportional
to:
� e��r
r; � =
s�2�E�h2
: (4)
where E is the single particle energy. For large but �nite r a polynomial in 1=r should be
included. Therefore, one expects the HFB neutron density to behave for large r, r > ro say,
as:
hqj�(r)jqi = �o(q)e�2�qr
r2
"1 +
bq
r+
cq
r2
#2; (5)
This form has been used to extrapolate the densities obtained in the multi-oscillator basis
for values of r beyond the values where the unphysical ledge begins. The parameters �q, bq
12
and cq were obtained by using a �2 �tting routine. �o(q) was chosen to match the small r
density pro�le at ro, where the tail was attached to the single oscillator calculation, typically
at about 7 fm. The o�-diagonal terms in the correlated densities (3) could then be computed
from:
hq0j�(r)jqi / e�(�q0+�q)r
r2
"1 +
bq0
r+cq0
r2
# "1 +
bq
r+
cq
r2
#: (6)
There is a great deal of freedom in the choice of the tail parameters and hence we were able
to obtain various di�erent parameter sets, depending on how we chose to �x things. The
�q parameters were set to be between 0:10 ! 0:15, which corresponds to energies of about
�0:21! �0:47 MeV.
With these �tted tails Fig. 8 is obtained. Di�erent sets of �tted tail parameters are
labeled by a, b, c, d, e. They give slightly but not very di�erent values for the extrapolated
density above 10�6. Note that here the tails extend further out than for the unconstrained
HFB case.
Let us now turn to result concerning rms radii. In Table I the rms radii for the various
densities of �gs. 4, 5 and 6 are shown. The correlations change the calculated rms radii for
the 9Li core by only about 2%, which is consistent with the previous observation that corre-
lations have a negligible e�ect upon the core. In this change the tail correction made above
plays almost no role. As to the di�erences in the rms radii between the 9Li nucleus and the
values calculated for the 9Li core, they are about 5% for both the protons and the neutrons.
This e�ect upon the rms radius appears to be small, but the comparison of the density dis-
tributions made in the previous section yields a more pronounced di�erence. One may say
13
that the 9Li system is slightly in ated when two extra neutron are added. This comparison
assumes that the 9Li nucleus ground state wave function is not signi�cantly in uenced by
correlations, which certainly is reasonable in view of the insensitivity of correlations on the
9Li core. Again, the comparisons made for the nine-nucleon systems don't depend on the
tail correction introduced previously. One may conclude that in this many-body calculation
of 11Li there appears to exist a substantive 9Li core which notably di�ers from a free 9Li
nucleus. One expects that these di�erences may have a nonnegligible e�ect upon the mean
�eld felt by the two outer neutrons.
Going to the full 11Li rms radii, the value listed in Table I with correlations included,
but without tail correction (3.42 fm), clearly is overestimated. In Table II the rms radii
obtained without and with the various parametrized tail corrections are listed. The largest
hrrmsitot using a �tted tail is case a with 2:88 fm. In the calculations performed here, the
increase in rms radii due to correlations is between � 0:04 ! 0:08 fm with realistic tails.
This result is somewhat disappointing in view of the the amount of complexity put into the
GS wave-functions. The total and neutron rms radii, although not so far from consistent
with the experimental error bars, appear 0:2 fm smaller than the nominal experimentally
extracted values.
At this point, one may note that the present radius results probably represent lowest
values. In fact, when using the multi-oscillator basis the resulting potential energy surface
appears somewhat atter than the nineteen shell basis shown in Fig. 1. This indicates that
that a fully realistic calculation where density tails would be correctly described, would
also yield a much softer collective potential and therefore a stronger con�guration mixing.
14
Hence e�ects of correlations larger than those derived here, especially with respect to the
rms radius predictions, would certainly be found.
Finally let us mention that, since the present 11Li calculations were performed assuming
axial symmetry, we were able to calculate the electric quadrupole moments, both in the HFB
and correlated calculations. It should be realized that since the proton sector was calculated
with a blocking approximation, the ground state spin of the nucleus has been arti�cially set.
In Table III the calculated axial quadrupole moments are displayed for 11Li and 9Li. Since
the proton sector is not greatly a�ected by the correlations, the HFB and correlated GCM
calculations give approximately the same charge quadrupole moment. Both of these results
give remarkable agreement with the experimental result. The free 9Li HFB result is also
shown, where here the agreement is not nearly as close.
IV. CONCLUSIONS
A microscopic many-body calculation of 11Li is presented with correlations. Pairing
correlations are included in a constrained HFB calculation, while important long-range cor-
relations are incorporated using a GCM type formalism. The sole input into the calculation
is the well-tested, �nite-range D1S e�ective interaction.
It was found that long-range correlations play an important role in the description of
11Li. A distinct 9Li core was found which remained uncoupled to both pairing and long-
range correlations. The correlations were restricted almost exclusively to the sector occupied
by the two valence neutrons. This substantive core appears to support the use of a three-
15
body framework to study this nucleus, although several factors must be considered. The
Pauli principle acting between the 9Li core neutrons and the two valence neutrons must
be taken into account as seen in Fig. 6. Also, the 9Li core found here appears to have a
signi�cantly di�erent density pro�le than a free 9Li nucleus. It may be necessary for three-
body calculations to take into account the di�erent density shape of the 9Li core if more
accurate calculations of this type are desired than those presently in existence.
Di�culties in this many-body treatment of 11Li were encountered due to the use of
naturally localized harmonic oscillator bases. Even with the use of an extended multi-
oscillators dependences upon the basis were found. For a more complete and accurate
calculation it would be necessary to use a better basis representation, which can accomodate
a nucleus with a large extended tail. For example a collocation Basis spline basis in a very
large box would probably work well.
Acknowledgements: Appreciation is extended to J. Luis Egido for bringing our attention
to this problem and to M. Girod, V. Khodel and R. M. Thaler for valuable discussions. One
of the authors (CRC) would like to acknowledge the gracious hospitality of the Service de
Physique et Techniques Nucl�eaires at Bruy�eres-le-Chatel where this research was performed.
16
REFERENCES
[1] M. V. Zhukov, B. V. Danilin, D. V. Fedorov, J. M. Bang, I. J. Thompson and J. S. Vaa-
gen, Phys. Rep. 231 (1993), 151.
[2] J. M. Bang, I. J. Thompson, M. V. Zhukov, B. V. Danilin, D. V. Fedorov and J. S. Vaa-
gen, Niels Bohr Institute preprint # NBI-91-31; J. M. Bang and I. J. Thompson, Phys.
Lett. B 279 (1992) 201.
[3] Y. Tosaka and Y. Suzuki, Nucl. Phys. A512 (1990) 46.
[4] A. S. Jensen and K. Riisager, Nucl. Phys. A537 (1992) 45.
[5] L. Johannsen, A. S. Jensen and P. G. Hansen, Phys. Lett. B 244 (1990) 357. P. G.
Hansen and B. Jonson, Europhys. Lett. 4 (1987) 409.
[6] A. C. Hayes, Phys. Lett. B 254 (1991) 15.
[7] I. Tanihata, T. Kobayashi, O. Yamakawa, S. Shimoura, K. Ekuni, K. Sugimoto, N. Taka-
hashi and T. Shimoda, Phys. Lett. B 206 (1988) 592.
[8] R. Anne et. al., Phys. Lett. B 250 (1990) 19; K. Riisager et. al., Nucl. Phys. A540
(1992) 365.
[9] G. F. Bertsch, B. A. Brown and H. Sagawa, Phys. Rev. C 39 (1989) 1154.
[10] H. Sagawa, Phys. Lett. B 286, 7 (1992).
[11] Z. Y. Zhu, W. Q. Shen, Y. H. Cai, and Y. G. Ma, Phys. Lett. B 328, 1 (1994).
[12] T. Hoshino, H. Sagawa and A. Arima, Nucl. Phys. A506 (1990) 271.
17
[13] J.-F. Berger, J.-Decharg�e and D. Gogny, Proceedings to the \Workshop on Nuclear
Structure Models", Oak Ridge, TN, March 16-25, 1992.
[14] J. Decharg�e and D. Gogny, Phys. Rev. C 21 (1980), 1568.
[15] P. Ring and P. Schuck, The Nuclear Many-Body Problem (New York: Springer-
Verlag, 1980).
[16] C. R. Chinn, J.-F. Berger, D. Gogny and M. S. Weiss, Phys. Rev. C 45 (1992) 1700.
[17] E. Arnold, J. Bonn, A. Klein, R. Neugart, M. Neuroth, E.W. Otten, P. Lievens, H. Reich,
W. Widdra and ISOLDE Collaboration, Phys. Lett. B 281, 16 (1992).
18
FIGURES
FIG. 1. The Total Hartree-Fock-Bogolyubov energy as a function of the constraint variable, hr2i
for 11Li and 9Li is shown. The 9Li resulted is taken from a spherically symmetric HFB calculation.
FIG. 2. In the upper panel the separate rms radii for the proton and neutron distributions
is plotted for the constrained HFB calculation as a function of the square root of the constraint
variable, the rms radii of the total matter distribution. The lower panel plots in a similar fashion
the BCS pairing energies for the neutron and proton sectors.
FIG. 3. The energies for some of the lowest single particle states are plotted for the protons
and neutrons in the upper and lower panels, respectively, for the constrained HFB calculation as
a function of hr2i.
FIG. 4. The neutron radial densities for the unconstrained HFB calculation are shown as a
function of r. The solid and short-dashed curves correspond to the calculated neutron densities
of 11Li and 9Li, respectively. The long-dashed curve represents the neutron density for the 9Li
core projected from 11Li. The lower panel is a logarithmic plot of the upper panel with the x-axis
extended.
FIG. 5. The proton radial density pro�les are shown as a function of r. The solid and
short-dashed curves correspond to the unconstrained HFB calculation for 11Li and 9Li, respec-
tively. The long-dashed curve is the result from the correlated GCM-type calculation. The lower
panel is a logarithmic plot of the upper panel with the x-axis extended.
19
FIG. 6. The neutron radial density pro�les are shown as a function of r. The solid and
short-dashed curves correspond to the unconstrained HFB calculation of 11Li and the projected 9Li
core of 11Li, respectively. The long-dashed and dot-dashed curves are the result from the correlated
GCM-type calculation for 11Li and the projected 9Li core of 11Li, respectively. The shaded, fat
long-dashed curve represents the di�erence between the long-dashed and the dot-dashed curves,
thus corresponding to the two outer valence neutrons. The lower panel is a logarithmic plot of the
upper panel with the x-axis extended.
FIG. 7. The neutron radial densities are shown for the constrained HFB calculation in a loga-
rithmic plot as a function of r. The solid and long-dashed curves correspond to calculations using
a 19 shell harmonic oscillator basis, where q is constrained to 3:42 and 4:02 fm2, respectively. The
short-dashed and dot-dashed curves use three concentric eight shell harmonic oscillators as a basis,
where q is constrained to 3:42 and 4:02 fm2, respectively.
FIG. 8. The collective neutron radial densities from the GCM-type calculation are shown,
where the arti�cial tails of various listed categories were used for the densities calculated in the
constrained HFB calculations.
20
Type of Calculation hrtotali [fm] hrprotoni [fm] hrneutroni [fm]
11Li nucleus, unconstrained HFB 2.80 2.30 2.97
11Li nucleus with correlations 3.42 2.31 3.75
9Li core projected from
11Li in unconstrained HFB 2.55 2.30 2.67
9Li core projected from
11Li with correlations 2.59 2.31 2.72
9Li nucleus, unconstrained HFB 2.47 2.24 2.58
11Li , experiment [7] 3.12(.16) 2.88(.11) 3.21(.17)
9Li , experiment 2.32(.02) 2.18(.02) 2.39(.02)
Table I: The calculated rms radii for the neutron, proton and total matter distributions of
11Li, 9Li and the 9Li core projected from 11Li calculation are listed. The results from
the pure mean �eld HFB and the correlated calculations are used.
21
With �tted tail hrtoti hrproti hrneuti
11Li nucleus, HFB 2.80 2.30 2.96
11Li w/correlations 3.42 2.31 3.75
11Li Case a 2.88 2.31 3.07
11Li Case b 2.84 2.31 3.02
11Li Case c 2.85 2.31 3.03
11Li Case d 2.86 2.31 3.04
11Li Case e 2.87 2.31 3.05
Table II: The calculated rms radii in fm for the neutron, proton and total matter distributions
of 11Li including correlations are tabulated using various �tted asymptotic tails.
22
Type of Calculation hQproton20 i [mB] hQneutron
20i [mB]
11Li unconstrained HFB -31.20 -11.08
Correlated GCM 11Li -31.13 2.87
9Li core projected from
11Li unconstrained HFB -31.20 -54.94
9Li core projected from
correlated GCM 11Li -31.13 -53.76
Free 9Li, HFB, D1S -43.46 -96.24
11Li, experiment [17] -31.2 (4.5)
9Li, experiment [17] -27.4 (1.0)
Table III: The calculated axial quadrupole moments in millibarns for the neutron, proton and
total matter distributions of 11Li, 9Li and the 9Li core projected from 11Li calculation
are listed. The results from the axial HFB and GCM calculations are used.
23